﻿308 Mr. Baly and Dr. Ramsay on the Relations of Pressure, 



contact with it. In the empty, or comparatively empty part 

 of the gauge, the capillary depression of the mercury is 

 almost certainly not so great as in that limb exposed to gas ; 

 and the alteration of the pressure of the gas doubtless also 

 alters the capillary depression. It would be worth while 

 making this the subject of special experiments. 



2. Expansion of Gases at Constant Pressure, on Rise 

 of Temperature. 



The last paper falling under our notice is one by G. Me- 

 lander, published in 1892 (Wied. Ann. xlvii. p. 136). It 

 deals with the thermal expansion of gases under reduced 

 pressure. He concludes that the coefficient of expansion of 

 air decreases from 0003666 at 752 millim. to 0*0036594 at 

 232 millim., and again increases, with reduction of pressure, 

 until at Q'6 millim. it amounts to 0*0037627 ; the turning- 

 point of carbon dioxide is 76*2 millim. ; with hydrogen there 

 is no change in direction, but the coefficient increases from 

 0*0036501 at 764*5 millim. to 0*0037002 at 9*3 millim. 

 These conclusions are opposed to those of Siljestrom and 

 Mendeleeff ; for if the pv increases, as stated by these 

 authors, it is to be expected that the coefficient of expansion 

 should decrease with decrease of pressure. 



Before concluding this historical sketch, we feel it neces- 

 sary to point out that it is exceedingly unlikely that these 

 experimenters were dealing with pure gases. It is true that 

 their apparatus were filled with comparatively pure gases ; 

 but they do not seem to have realized the fact that on reduc- 

 tion of pressure, gas different from that with which they 

 were dealing comes off the walls of the containing vessel. 

 When this is of glass, a very considerable amount of carbon 

 dioxide is always evolved. At high pressures, of course, this 

 would make a small difference, probably inappreciable ; but 

 as the pressure becomes lower, a gas like hydrogen will be 

 partly removed, and its place taken by carbonic anhydride. 

 One can never be certain of the purity of the gas measured 

 unless its vacuum-tube spectrum is free from that of other 

 gases. 



Another source of error, small in itself, but possibly im- 

 portant when the small deviations which the experimenters 

 are endeavouring to measure are taken into consideration, is of 

 the nature of that pointed out by Lord Rayleigh as applying to 

 Regnault's determination of the comparative weights of hy- 

 drogen, oxygen, and nitrogen. It is that the volume of a glass 

 (and presumably of an iron) vessel is considerably diminished 



